Highlights
Denial of Service: 4 prior fixes. Scrutinize any change in this area.
gensim/models/word2vec_corpusfile.pyx: most-fixed (2 issues). Treat as high-risk during review.
0 high-severity fixes in this history; regressions here are high-impact.
Recurring patterns
The bug types that recur here, drawn from past fixes, not open vulnerabilities.
Denial of Service: Integer overflows during word counting and metrics evaluation in corpus file-based training can trigger negative values or unexpected program termination when processing extremely large datasets.
Denial of Service: Parsing large MatrixMarket term-document matrices can trigger integer overflow exceptions or crashes if document or term counts exceed 32-bit bounds during conversions like __Pyx_PyInt_As_int.
Denial of Service: Concurrency issues in the LDA distributed training worker thread dispatching can cause infinite blocking loops and thread leaks, starving resources.